When starting any supervised learning problem, the first step is to load and prepare the data. We are going to start by loading the MNIST Fashion dataset[3], a collection of small, grayscale images showing different items of clothing. Our job is to build a system that can recognize what is in each image; that is, does it contain a dress, a shoe, a coat, and so on?
First, we need to download the dataset by running the download-fashion-mnist.sh script in the code repository. Then, we will load it into Go:
import (
"fmt"
mnist "github.com/petar/GoMNIST"
"github.com/kniren/gota/dataframe"
"github.com/kniren/gota/series"
"math/rand"
"github.com/cdipaolo/goml/linear"
"github.com/cdipaolo/goml/base"
"image"
"bytes"
"math"
"github.com...